Fluctuations under turbulent flows: Enhanced methods for separation of propagative wavenumbers from wall pressure dataset

Abstract

This work is a part of a more general study on automobile interior noise due to acoustic and aerodynamic wall-pressure fluctuations. Using experimental data of wall-pressure fluctuations measured with a microphone array beneath several kinds of flows, a wave-number analysis based on recently developed signal processing methods -- the spatial Empirical Mode Decomposition (sEMD) and the Ensemble Empirical Mode Decomposition (E-EMD) -- is carried out, in order to separate acoustic and aerodynamic pressure fluctuations. In opposition with an existing classical method previously used, based on the spatial correlogram, these methods do not require a stationary uniform flow. A turbulent boundary layer on a flat plate and a detached/reattached flow downstream three different forward-facing steps are tested, with flow velocities from 0 to 40 m.s-1, with or without a well-controlled artificial acoustic source. The sEMD method is first used as a wavenumber filter, and is shown to improve the detection of acoustic fluctuations of about 5 dB on classical wavenumber (k,f) representations. The E-EMD method is developed in order to decompose the (x,t) representation to make possible the separation of the acoustic and aerodynamic components, regardless of the stationarity or the uniformity of the flow.